30 second clips
One continuous take up to 30 seconds, roughly double the 15 to 20 seconds you get from most rivals, and no stitching needed.
Seedance 2.5 is ByteDance's newest AI video model. It makes one continuous take up to 30 seconds long, generates the sound in the same pass, and takes image, video and audio references to steer the shot. It is live on avots.ai right now, so sign up and make a clip in a couple of minutes.
ByteDance showed Seedance 2.5 in June 2026. Here is what changes versus Seedance 2.0, based on the announcement.
One continuous take up to 30 seconds, roughly double the 15 to 20 seconds you get from most rivals, and no stitching needed.
ByteDance announced 4K for its own app. The providers serving Seedance 2.5 today expose 480p and 720p, and that is what avots renders. If 4K opens up it lands on the same balance.
Steer a shot with up to 30 photos, 10 clips and 10 audio files, 50 in total, and point at them in the prompt as @Image1, @Video1, @Audio1. Seedance 2.0 stops at 12.
Audio is made together with the picture, so on screen action and sound line up without extra editing.
ByteDance claims about 20 percent tighter prompt adherence, which means fewer tries before you get a usable shot.
You pay for the seconds you render: about 2,200 tokens for a 5 second clip at 480p. The card shows the exact cost before anything is charged.
No separate account, no extra API key, no subscription. Create one avots account and the same balance works everywhere.
Open app.avots.ai, pick the model, write a prompt, download the result. Works on desktop and mobile.
Create from your phone in @AvotsAIbot. Same models, same balance.
Call it from Claude, Cursor or Cline through the avots MCP server.
Script it into your own app with the OpenAI compatible API and one key.
| Seedance 2.5 | Veo 3.1 | Sora 2 Pro | Kling v3.0 Pro | |
|---|---|---|---|---|
| Max clip length | up to 30s | about 8s | up to 15 to 25s | about 10s |
| Resolution | 480p and 720p | up to 1080p | up to 1080p | up to 1080p |
| Sound | yes | yes | yes | yes on Pro |
| Reference inputs | up to 50 | image | image | image plus elements |
| On avots.ai | live now | live now | live now | live now |
Competitor numbers are rough and depend on tier and settings. On avots.ai every model runs from one balance, so you can put them side by side on the same prompt.
Seedance 2.5 is on avots.ai, with Veo, Sora, Kling and Nano Banana on the same balance. Pay for the clips you make, no subscription and no auto renewal.
Sign up and create ✦ByteDance video with native audio and reference to video. Live now.
Google Veo, realistic AI video with sound. Live now.
OpenAI Sora, strong physics and longer shots. Live now.
Google Gemini image model for photoreal images and edits. Live now.
FLUX 1.1 Pro and FLUX.2 for photoreal images. Live now.
Kuaishou's Kling video model, strong motion. Live now.
And many other models. avots.ai runs 40+ AI models for chat, image, video and audio on one balance. See the full list on the Tools page, or open the app and start creating.